Introduction

Microalgae, at the base of the trophic chain, are model photosynthetic organisms in stress study [1]. These low-cost, rapid response and stress-sensitive organisms are used, due to their short term of generation, in most European toxicity standard tests. They are therefore considered as bio-indicators of pollution [2]. In this study, we shall focus on Chlamydomonas reinhardtii, a unicellular biflagellate organism with rapid growth. This specie is an ecotoxicological model ubiquitously present in aquatic environments [3]. In ecotoxicology, survey of macromolecular composition variation (proteins, nucleic acids, lipids, carbohydrates) highlights microalgal physiological variation after an environmental change involving stress. Several modifications existing naturally in the environment such as lack of light (the most limiting factor according to Mata et al. [4]), variation of temperature, nutrient element depletion, changes in CO2 amount, or salinity are stressing conditions for microalgae. The stresses induce different types of responses. Salinity adjustment can increase lipid production of microalgae and yield enhancement of astaxanthin and β-carotene pigment production [5]. Enhancement of protective pigments is also measured when light intensity rises [5]. Photosynthetic system disruptions of Chlamydomonas reinhardtii after high light exposure (1000, 1500 and 2000 μmol photons.m−2.s−1) during 1 h were described by Nama et al. [6]. Indeed, variations of illumination hinder light capture and cause decreased photosynthetic yield (less efficient transmission to photosynthetic reaction centre and damaged protein profile) [6]. Furthermore, lack of phosphate (P) in the medium has been shown to result in photosynthesis repression [7]. Stress towards microalgae may also induce macromolecular production. In nitrogen restriction (N), lipid accumulation has been observed for five strains of Chlorella [8]. Nitrogen depletion also leads to enhancement of secondary carotenoids and competitive attribution of carbon, which is additionally required for protein synthesis [9]. Finally, Chiu et al. [10] showed a biomass rise and lipid accumulation accompanying CO2 increase in Nannochloropsis oculata culture.

All of these effects, via compound shortage or accumulation, alter microalgal chemical content. Monitoring of their composition is consequently a good means of evaluating microalgae stress and is frequently carried out by biochemical assays such as the BCA (proteins) and Dubois (carbohydrates) methods [11, 12]. Among the numerous techniques used to structural characterization of various biological systems [13], recently, vibrational spectroscopy Fourier transform infrared (FTIR) has been described as a reliable and sensitive tool for rapid measurement of some phytoplankton species’ macromolecular composition [14, 15]. For example, infrared microspectroscopy has monitored the modification of metabolites’ spatial distribution in Micrasteria hardyi after exposure to an active pharmaceutical ingredient [16].

In cases of nutrient starvation, Driver et al. [17] have shown that infrared spectroscopy follows impacts on microalgae despite its limits in terms of microalgae specie identification. By this technique, biomass composition changes after limitations of nutrients (N and P) have been studied over recent years [18,19,20,21]. Stehfest et al. [21] followed the physiological stages of cyanobacteria and bacillariophyceae by measuring their photosynthetic efficiency (short-term effect) and by recording their FTIR signatures. After 2 weeks of nutrient starvation (N or P), using FTIR data, they observed a highly specie-specific metabolic answer and variation in the area ratios of different biochemical groups (proteins, lipids, carbohydrates and silicate). Ratio variations were first used by Jakob et al. [22] to carry out quantitative tests of cellular macromolecules (proteins, carbohydrates and total lipids). Wagner et al. [15] suggested that infrared spectroscopy is well adapted for macromolecular quantification using reference spectra (bovin serum albumin, glucose, tripalmitate glycerol) and PLS (partial least squares) algorithm rather than biochemical methods. Infrared has also been used to study cellular carbon allocation according to phosphorus quota in Chlamydomonas reinhardtii [18] and Scenedesmus subspicatus [19]. In cases of nitrogen depletion, carbon allocation in marine microalgae, Isochrysis zhangjiangensis, has been monitored by determination of absorption band ratio [20].

This study aims to demonstrate the relevance of infrared absorption spectroscopy coupled with principal component analysis in the study of macromolecular composition variations, some of them quite small, in a model microalga. We first applied the procedure to controlled environmental changes, after which we investigated interaction with a potential disruptor, nanoplastics, which were recently described as present in aquatic media [23]. In the literature, polystyrene nanobeads (PS) are frequently chosen as a model for nanoplastics [24]. Micro or nanoplastic effects on microalgae have been studied, for example, by Sjollema et al. [25], who reported reduced growth of Dunaliella tertiolecta exposed to 50 nm PS nanoparticles. In addition, Bhattacharya et al. [26] used SEM to observe an aggregation of PS nanobeads on the flagella of Scenedesmus. Several authors have also demonstrated that interaction mechanisms of PS NP with microalgae are different according to surface charge [26,27,28]. D. tertiolecta, in interaction with PS-COOH NP (up to 50 μg/mL), does not seem to affect growth (aggregation phenomenon), contrary to NH2-coated PS NP, which inhibits growth (EC50 = 12.97 μg/mL) [29]. Hazardous effects have been reported in cases of positively charged PS due to strong adsorption on cell wall, leading to ROS production [26, 28, 30]. Finally, these studies have not clearly highlighted an effect of polystyrene NP on microalgae. In this work, the potential impact of nanoplastics on microalgae chemical composition, which was not studied before, was investigated by infrared spectroscopy and compared to the effects measured under other controlled environmental stresses.

Materials and methods

Study model: Chlamydomonas reinhardtii

The wild-type Chlamydomonas reinhardtii strain 2935 (Chlamydomonas Resource Center, [31]) was cultured in 250 mL Erlenmeyer flasks containing 100 mL of tris acetate phosphate (TAP) medium [32] at 20 °C. Erlenmeyer flasks were placed under 122 μmol photons.m−2.s−1 of irradiance. Illumination was brought about with white-cold fluorescent tubes (Philips TLD 36 W) and followed a 14/10 h light/dark cycle. Cultures were maintained in constant suspension by agitation at 100 rpm (VWR, Model 5000).

Controlled stress conditions

Microalgae were subjected to different light intensities and to reduced of nitrogen intake during 48 h. The control group was illuminated with an intensity of 122 μmol photons.m−2.s−1 whereas the low light group was illuminated with only 90 μmol photons.m−2.s−1 and the high light group with 350 μmol photons.m−2.s−1. Another group was placed in the dark, using aluminium paper recovery. Nitrogen intake was reduced by 25% compared to the control group.

Model nanoplastics

Microalgae (1 × 106 cells/mL) were exposed to polystyrene (PS) nanoparticles for 72 h. PS nanoparticles of 100 nm in diameter were purchased from Interchim. According to the supplier’s instructions, PS nanoparticles were stabilized with anionic surfactant in aqueous solution at 1% in concentration (stock solution). They were diluted at 0.01% in ultrapure water, and 100 μL was added to 20 mL of algae solution to obtain a final concentration of 1 mg/L corresponding to 9.15 × 107 NP/microalgae. Zeta potential and size were controlled with a Zetasizer Nano ZS device (Malvern). In ultrapure water at 0.001% of concentration, PS nanoparticles were monitored at an average size of 240 nm and a zeta potential of − 11 ± 1 mV.

Interaction visualization by confocal microscopy

Microalgae are naturally fluorescent and PS nanoparticles were marked with Nile red. Analyses were performed on Zeiss LSM800 microscope (Carl Zeiss Microscopy GmbH, Jena, Germany). Microalgae and PS nanoparticles fluorescence’s were observed after 2 h of interaction. One thousand microlitres of the sample was placed in cavity slides, hermetically sealed with a cover slip, and observed at × 25 objective under excitation by two lasers at wavelengths of 488 and 561 nm with 5% of intensity. The light emitted by the sample was collected by two detection channels, the first with wavelengths of 583 to 700 nm (master gain: 530 V) and the second with wavelengths of 550 to 583 nm (master gain: 594 V).

Cytotoxicity measurement

To quantify membrane damage, the LDH test was established in triplicate on one microalgae solution after contact with PS nanoparticles for 1 h and 72 h. In the cell, lactate dehydrogenase enzyme (LDH) catalysed oxidation reaction of lactate to pyruvate. This oxidation reaction is directly linked to the reduced co-reaction of NAD+ (oxidized form of nicotinamide adenine dinucleotide) in NADH (reduced form). The test measured NAD+ quantity reduced by absorbance on a UV-visible spectrophotometer. The protocol was the one described by Pakrashi et al. [33].

Spectroscopic monitoring

Infrared spectra were acquired to monitor microalgae composition in transmission using a FTIR-microscope Spotlight 200i, Perkin-Elmer from 4000 to 500 cm−1 with an accumulation of 32 scans and resolution of 4 cm−1. Five replicates of 10 μL of the microalgae solution were deposited on a CaF2 surface and one spectrum was acquired for each droplet. The characteristic bands of interesting compounds in microalgae such as lipids (1742 cm−1, C=O), proteins (1650 cm−1, Amide I), nucleic acids (1255 cm−1, P=O) and carbohydrates (1045 cm−1, C–O) [18, 20] were measured in intensities, and ratios were calculated from raw data. After application of a linear baseline treatment to all spectra, average infrared spectra of the studied area (950–1900 cm−1) with standard deviation were determined with Excel 2007 software, following which preprocessing using the first derivative Savitsky-Golay (D1SG) function was applied to the spectra to smooth the data, followed in turn by standard normal variate (SNV) treatment with zero mean. The D1SG+SNV method was proposed by Rammal [34] as one of the preprocessings most adapted to spectroscopic dataset prior to chemometric analysis. Principal component analysis (PCA) was then performed on the dataset after preprocessing with Unscrambler software.

Results and discussion

Technique validation on controlled stress conditions

The choice of relevant spectral bands to monitor changes in microalgae physiology has previously been discussed in the literature. Arguments for entire spectral study were set forth by Giordano and al. [35] and contributed to analysis of spectra sectors identified by Driver et al. [17]. The most commonly chosen part of the spectra is the one containing most of the bands of interest, between 950 and 1900 cm−1 [15, 17]. Here, the area chosen for the following analyses was restricted to 950 to 1800 cm−1, allowing elimination of the O–H chemical bond (3200–3500 cm−1), which is linked to water quantity in samples.

The average infrared spectra of microalgae under controlled environmental changes are presented in Fig. 1a and the ratios between bands of interest are plotted in Fig. 1b. Ratio of proteins on carbohydrates bands (I1650/I1045) increased for dark and low light conditions. In accordance with Jebsen et al. [14], protein rather than carbohydrate production is promoted in (partial or total) light reduction. On the contrary, the I1650/I1045 ratio decreased for nitrogen stress and high light condition compared to control. Hu [36] observed a rise of carbohydrates with high light intensity. In cases of nitrogen depletion, this observation is consistent with the study of Stehfest et al. [18], who showed that under moderate N depletion, an increase of carbohydrates initially occurs. Concerning lipids, it has been shown in the literature that under the high stress caused by nutriment depletion, microalgae produce lipids [18]. This is confirmed here as the ratio I1742/I1650 and I1742/I1045 increased only with N depletion. In addition, a drop in lipid/protein ratio (I1742/I1650) is observed for dark and low light conditions. Patel et al. [16] have also reported this decrease, leading to impacts on the lipid bilayer of Micrasterias hardyi that have appeared after interaction with active pharmaceutical ingredients. For high light condition, the change of lipid content compared to proteins remained moderate. Enhancement of light also leads to a decrease of lipids compared to carbohydrates (I1742/I1045), and this result is confirmed by Benavente-valdés et al. [37].

Fig. 1
figure 1

Microalgae infrared spectra (48 h, 950–1800 cm−1) under controlled environmental changes with standard deviation (control (full blue): 122 μmol photons.m−2.s−1, N- (green tile): nitrogen depletion, dark (black straight line): without light, high light (red diamond): 350 μmol photons.m−2.s−1, low light (orange inclined line): 90 μmol photons.m−2.s−1) (a) after linear baseline (LB); (b) ratio on principal characteristic bands of LB spectra under know environmental changes with standard deviation: lipids (1742 cm−1, C=O), proteins (1650 cm−1, amide I), nucleic acid (1255 cm−1, P=O), carbohydrates (1045 cm−1, C–O)

Full size image

Furthermore, results concerning nitrogen deficiency are in line with the conclusions of existing studies indicating an accumulation of lipids and carbohydrates [38]. For example, Meng et al. [20] used FTIR to measure protein carbon allocation deflection for the benefit of carbohydrates and neutral lipids in Isochrysis zhangjiangensis. Finally, low light condition presents increased nucleic acid/carbohydrates ratio (I1255/I1045), conversely, with enhancement of light the ratio declines. These results underline the redirecting of microalgae metabolism and change in energy allocation: storage compounds (proteins, carbohydrates and lipids) are preferred under stress conditions [14, 18, 38].

Spectra analysis by band ratio, as previously described, is informative but is not easily applicable to a sizable number of spectra. Spectral statistical treatment is needed and PCA was chosen. On these infrared spectra dataset, PCA (inputs variables: wavenumber and absorbance dataset; outputs: PCA individual plot on the two first PC and X loadings; total PC number: 3) is carried out after application of D1SG+SNV preprocessing (Figs. 2 and 3). This preprocessing procedure is optimized to this dataset in accordance with literature [34].

Fig. 2
figure 2

Microalgae infrared spectra after D1SG preprocessing (first derivative of Savitzky-Golay) + SNV (48 h, 950–1800 cm−1) under controlled environmental changes with standard deviation (control (blue): 122 μmol photons.m−2.s−1, N- (green): nitrogen depletion, dark (black): without light, high light (red): 350 μmol photons.m−2.s−1, low light (orange): 90 μmol photons.m−2.s−1

Full size image
Fig. 3
figure 3

a PCA on D1SG+SNV microalgae spectra (48 h, 950–1800 cm−1). b Loadings principal component 1 (PC1, blue). c Loadings PC2 (red) and d loadings PC3 (green). Control (blue): 122 μmol photons.m−2.s−1, N- (green): nitrogen depletion, dark (black): without light, high light (red): 350 μmol photons.m−2.s−1, low light (orange): 90 μmol photons.m−2.s−1

Full size image

PCA is used to cluster spectra by similitude and helps to distinguish different conditions from one another. In our case, dark and low light assays on the one hand and high light and nitrogen depletion on the other hand are distinctly divided into two large groups on the PC1 axis, with the control group at the centre. To explain these clustering, loading plots were inspected. PC1, which accounts for 54% of spectral variation, presents fluctuations in lipids (C=O, 1740 cm−1), proteins (C=O, 1650 cm−1; N–H, 1545 cm−1) and carbohydrate peaks (v C-O-X, 1150 cm−1; 1078 cm−1; 1022 cm−1) in accordance with the results described by Dean et al. (2008). Regarding PC2 (25% of explained variance), dark and low light conditions are distinctly divided into two clusters of C. reinhardtii spectra. The bands on this component that differ are connected to carbohydrates (1150; 1080; 1022 cm−1). PC3, which accounts for 11% of spectral variation, also presents changes on carbohydrate bands (1150; 1045; 1022 cm−1). Carbohydrates bands can be assigned to glycosidic bond C–O–C and C6-O6-C1 respectively for 1150 and 1080 cm−1 peaks whereas bands at 1045 and 1022 cm−1 inform about carbohydrates deformation [39, 40].

This experiment validates the hypothesis that controlled stresses due to environmental modifications lead to infrared spectrum modification of the microalgae. Changes are notably observed on protein and carbohydrate bands. Ratio of protein to carbohydrate (I1650/I1045) content in microalgae appears to fluctuate according to light intensity or nutrient depletion, as shown in the literature concerning nutrient stress [21, 22]. These results give complementary information regarding to the influence of light on the biochemical composition of microalgae [5].

In addition, the macromolecular compound variations according to culture conditions highlighted by raw spectra ratio study have also been reported by principal component analysis (PCA), which presents two advantages compared to ratios, namely the processing of statistics and detailed analysis. Major variations are observed for protein and carbohydrate bands. D1SG+SNV preprocessing is adapted to this dataset as more than 90% of spectral variance is explained by PCA.

Application of technique to a study case: presence of nanoplastics

Infrared spectra obtained after microalgae interaction with PS nanoparticles are presented in Fig. 3. Contrarily to the modifications observed under other controlled stress conditions, lipid/protein (I1742/I1650) and lipid/carbohydrate(I1742/I1045) ratios on the spectral bands (Fig. 4c) did not display large variation compared to the control. Lipid content proportions seem stable and microalgae do not appear to store energy under this form in the presence of PS NP. This result indicates a marginal effect of PS NP on microalgae lipids composition. The changes due to PS NP presence at this concentration were not comparable with the nutrient depletion effect [19, 21] nor with metal nanoparticle presence [41] which led to an accumulation of lipids. The concentration of 1 mg PS NP /L here is in coherence with previous studies where concentrations ranged between 0.1 and 50 mg/L [27, 29, 30]. The concentration was chosen in the lower range while allowing a sufficient quantity of NP to interact with microalgae.

Fig. 4
figure 4

Microalgae infrared spectra (72 h, 950–1800 cm−1) in contact with PS nanoparticles (1 mg/L, purple dotted line) and control (blue solid line): a after LB and b after D1SG+SNV. c Ratio on principal characteristic bands of infrared spectra in contact with PS nanoparticles (purple) and control (blue): lipids (1742 cm−1, C=O), proteins (1650 cm−1, amide I), nucleic acid (1255 cm−1, P=O), carbohydrates (1045 cm−1, C–O)

Full size image

In addition, an increased protein/carbohydrate (I1650/I1045) ratio and a decreased nucleic acid/carbohydrate ratio (I1255/I1045) in the presence of PS NP were observed. As in the case of lipids, carbohydrates did not seem to contribute to this type of energy storage as carbohydrate bands in the presence of PS NP appeared less intense than the control. As concerns proteins, under weak illumination conditions, we observed increased protein production, as also reported by Jebsen et al. [14] under light limited conditions (50 μmol photons.m−2.s−1). Similarly, protein synthesis was already described in Scenedesmus in iron limitation conditions [38, 42]. According to Fig. 3a, microalgae proteins content increased in the presence of PS NP, suggesting that microalgae react to the presence of PS NP. The increase of protein transferred from assimilated carbon may be used to promote microalgae growth as already described by Wagner et al. [43]. Finally, as in high light condition, the nucleic acid content of microalgae in the presence of PS NP was very weak compared to the control group.

As shown on Fig. 5, PCA presents a differentiation of the two conditions according to PC1, which explains 61% of variance. The first component is mainly influenced by band variations at 1650, 1545 (proteins), 1411 and 1052 (carbohydrates) cm−1. PC2 outline fluctuation on 1148, 1078, 1067 and 1019 cm−1 peaks are related to carbohydrates (Fig. 5b). Furthermore, PC3 represents the background noise of spectra (8%, Fig. 5c). These fluctuations are similar to the environmental changes described in the literature and remain relevant to moderate stress conditions [17, 44]. For example, Driver et al. [17] presented variations of the same protein and carbohydrate bands under low nutrients conditions. Due to this moderate stress, the control condition seems more variable than that subjected to more severe stress.

Fig. 5
figure 5

a PCA on D1SG+SNV microalgae spectra (72 h, 950–1800 cm−1) in contact with PS nanoparticles (1 mg/L, PS 2935): control: blue dot; PS 2935: purple square. b loading principal component 1 (PC1) and c loading PC2

Full size image

In the same way, spectra variability of microalgae in interaction with PS nanoparticles appears here also more important than differences with the control. Indeed, one culture batch appears very close to the control whereas the two other cultures batch are well discriminated on the PC1 axis. Such variability depending on the Erlenmeyer flask has been already observed and described in other studies [45] such as Driver et al. [17], where infrared spectra of C. reinhardtii and C. concordia in controlled conditions presented the same type of variability. Moreover, variability can be explained by a differential distribution of nanoparticles in the medium, leading to microalgae differentially impacted by PS NP. Another hypothesis could be that the modifications created by the interaction are reversible. In any case, it can be suggested from these results that nanoplastics in the environment surrounding microalgae have at least a short-term and moderate impact on microalgae biochemical composition. Infrared microspectroscopy allowed here to obtain information impact on microalgae biochemical composition after an interaction of 72 h which was no studied before. However, compared to other environmental changes, the variability between samples seems to be more important and the impact appears moderate so it would be interesting to determine whether this effect is observed in the long term, especially with regard to transgenerational effects [2].

Interaction with nanoplastics: visualization and membrane damage quantification

After demonstrating microalgae metabolic flux changes (I1650/I1045, protein/carbohydrate ratio and I1255/I1045, nucleic acid/carbohydrate ratio), by infrared microspectroscopy, it is relevant to confirm the interaction between PS nanoparticles and microalgae. To achieve this, microalgae interaction with PS nanoparticles was initially visualized by confocal microscopy (Fig. 6). The PS nanoparticles seem adsorbed on C. reinhardtii surface.

Fig. 6
figure 6

Interactions visualization of C. reinhardtii PS nanoparticles with confocal microscope after 2 h of exposure (blue: microalgae structure; yellow: PS NP)

Full size image

Membrane damage was monitored by the LDH test. LDH enzymes were not supposed to be found in the culture medium in cases where the plasma membrane of microalgae is unaffected. When the membrane is impacted, LDH enzyme quantity in the medium is proportional to the quantity of membrane damaged during interaction. A significant increase of LDH enzyme percentage compared to the control has been measured in all samples, 11.2% after 1 h and 7.5% after 72 h of interaction with PS NP. These results suggest that plasma membrane integrity of C. reinhardtii appeared to be impacted in the presence of PS NP. This phenomenon was also shown by Pakrashi et al. [33] in Chlorella ellipsoids after interaction with aluminium oxide nanoparticles of 100 nm in comparable concentrations (0.05 μg/mL) with a rise of 13% after 72 h. Likewise, at concentrations ten times higher, ZnO presence of 50 nm also leads to membrane integrity loss in Chlorella vulgaris [46]. It was already described that nanoparticles in contact with microalgae can cause morphologic damage leading to the non-integrity of microalgae barriers [46, 47]. However, recovery is uneven with some areas extensively covered. In this case, problems of nutrient exchange between the microalgae and the external environment could occur. This phenomenon of interaction and toxicity has been shown in the Skeletonema costatum diatom [48]. With green microalgae, however, interaction with plastic could be positive in the medium term. In 2018, Mao et al. [49] showed that the interaction of the green microalgae Chlorella pyrenoidosa with polystyrene microplastic resulted in a thickening of the cell wall involved in its rehabilitation and growth recovery. Here, the percentage of membrane damage decreased over time, suggesting that the cell adopted repairing mechanisms, leading to protein production [50]. These results can be probably related to the intensity variation noted with the infrared spectra on protein bands and may also explain the variability in the infrared spectra that were observed here.

Conclusion

The effect of strong environmental stresses such as nutrient or light depletion on the biochemical composition of microalgae Chlamydomonas reinhardtii could be easily highlighted by infrared spectroscopy coupled to multivariate analysis such as PCA. This technique has the advantage of requiring a low quantity of microalgae and, contrary to standard biochemical assays, of providing a rapid response. PCA method was complementary to band ratio calculation in attempts to understand the major biochemical modifications involved in stress response. In this study, infrared microspectroscopy was used for the first time to evaluate the effects of PS NP on microalgae by monitoring its biochemical composition. Our results showed for C. reinhardtii that interaction with nanoplastics led to a moderate response compared to nutrient lack or light variations. Indeed, lipid content remained stable, only carbohydrates content slightly decreased together with an increase of the proteins content. However, PS NP presence could be assimilated to an environmental change as they induced in situ modifications in microalgae. Interaction between microalgae and PS NP was confirmed by confocal microscopy, where PS aggregates were seen as adsorbed on C. reinhardtii surface. Moreover, LDH test exhibited a loss of C. reinhardtii membrane integrity after 72 h of interaction with PS NP. The resulting chemical modifications in the membrane were probably related to the biochemical variations highlighted by infrared spectra analysis. Overall, this study shows that infrared coupled with appropriate chemometric analysis can be very helpful in the understanding of nanoparticles-microalgae interaction mechanisms.